Application of the EU AI Act in the Military, Defence and National Security Sectors
As the world’s first comprehensive AI regulation with direct effect across all member states, the EU’s AI Act (the “Act”) establishes a tiered, risk-based framework for AI systems and imposes increasingly stringent requirements regarding transparency, auditing, and oversight as the levels of risk increase. However, in a military, defence and national security context, the Act contains a critical carve-out.
The Carve Out
Article 2(3) carves out from the Act’s scope all AI systems placed on the market or put into service “exclusively for military, defence or national security purposes, regardless of the type of entity carrying out these activities”. Similarly, AI systems which are not placed on the market or put into service in the Union, but where the output is used in the Union exclusively for the purposes mentioned above, the Act does not apply. Understanding the Act’s relevance to the military, defence and national security sector (the “Defence Sector”), requires a consideration of the scope of the exemption itself.
Recital 24 of the Act explains that the exemption is justified because “national security remains the sole responsibility of Member States” under Article 4(2) TEU, and that for military AI applications “the more appropriate legal framework is public international law.”
Importantly the same Recital 24 qualifies: if an AI system developed for military, defence or national security purposes is subsequently used, temporarily or permanently, for other purposes — for example civilian, humanitarian, law enforcement or public security purposes — it falls within the Act’s scope, and the entity using it for such purposes must ensure the compliance of the AI system with the Act. Further AI systems placed on the market or put into service for an excluded purpose, namely military, defence or national security, and one or more non-excluded purposes, such as civilian purposes or law enforcement, fall within the scope of the Act and providers of those AI systems should ensure compliance with the Act, for all non-military use cases.
Therefore, what appears to be a straightforward blanket exemption, whereby declaring an AI system’s military purpose is sufficient to remove it from the Act’s reach, becomes more complicated where that AI system’s use changes or indeed where it is a dual-use system.
Further given the Court of Justice of the European Union’s consistent position that any derogation from EU law on grounds of national security must be interpreted strictly, reliance on Article 2(3) as a blanket exemption against the Act’s requirements is legally vulnerable.
The Dual-Use Problem
Perhaps the most significant and regulatory challenge for European AI companies working within the Defence Sector is determining whether any given system qualifies for the exemption given that today’s AI innovations in large part originate in the commercial sector before being adapted for use in the Defence Sector. The European Defence Agency’s May 2025 White Paper on Trustworthiness for AI in Defence (the “White Paper”) acknowledges this, noting that the distinction between civilian and military AI applications is structurally unclear and that civilian AI innovation cycles often set the pace for defence capabilities.
Filling any Governance Gap?
It is noteworthy that the White Paper introduces governance requirements for Defence Sector AI — including Requirement Identification, Mandatory Impact Analysis, and considerations around human-AI race conditions — that align closely with the Act’s high-risk framework, even though the Act itself does not apply to Defence Sector AI systems. NATO’s revised 2024 AI Strategy adopts a similar approach - endorsing six Principles of Responsible Use — Lawfulness, Responsibility and Accountability, Explainability and Traceability, Reliability, Governability, and Bias Mitigation — mirroring the Act’s obligations without creating binding obligations.
Conclusion
As Europe scales up its defence spending, AI is quickly becoming central to the EU’s vision of technological sovereignty. The application of the AI Act therefore comes into sharp focus. While the Act excludes from its scope AI systems used exclusively for military, defence and national-security purposes this should not be viewed as a blanket exemption. The availability of the exemption depends on the AI system being used exclusively for those purposes. Defence Sector companies should therefore assess whether each use genuinely falls within that exemption and monitor the evolving guidance from the European Commission and national competent authorities.
Our Aerospace, Defence & Security team is ready to advise businesses operating in this rapidly evolving area. Find out more about our expertise on McCann FitzGerald's Aerospace, Defence & Security page.
This document has been prepared by McCann FitzGerald LLP for general guidance only and should not be regarded as a substitute for professional advice. Such advice should always be taken before acting on any of the matters discussed.





Select how you would like to share using the options below